1. Field of the Invention
This invention relates to a data processing technique, and more particularly, to an improved data processing method for supporting communications among two or more nodes on a network in a distributed computer environment.
2. Description of the Related Art
There is a mobile agent technology in the art in which an internal condition is moved to a place which is provided in a server existing on a network in a distributed computer environment to form an instruction in the place where the internal condition is moved Reference is made to U.S. Pat. No. 5,603,021, issued Feb. 11, 1997, entitled “System and Method for Distributed Computation Based Upon the Movement, Execution, and Interaction of Processes In A Network”, by White et al., Japanese unexamined Patent Publication No.: 7-182174, and an article entitled “Latest Internet Technology, special edition of Nikkei Communication”, by Fumihiko Nishida, Susumu Fujiwara et al., pp. 104-117, Nikkei B P.
Mobile agents, operating within the above mentioned mobile agent technology, make a contact to other agents in the moved-to place, i.e. a mobile agent or a resident agent, and may be served as appropriate. A place is a location provided by a server existing on the network to which the agent is moved and which supports a contact among agents. Preferably, the server accommodates the differences between the hardware and between the platforms.
The mobile agent technology enables a mobile agent to act as a proxy of a human being in handling work such as, for example, dynamically adjusting a schedule of internal meetings in compliance with the schedule of attendees and the status of reservations of conference rooms, and also acquiring desired information distributed on the network.
In the prior art mobile agents, however, it can be difficult to realize a flexible, smooth information exchange and coordination among agents because the communications among agents are realized by a conventional method call (subroutine call) or by low level message passing. Further, it is often a requirement to define the processing of a stereotypical conversation in order to communicate by using an agent. This requirement can result in a load to a user of an agent. Furthermore, it can be difficult to communicate between different types of agents because no agent communication language is used to realize the communications among a variety of agents.
It can be appreciated that a “conversation” between human beings generally consists of steps in which:                1. A speaker prepares a content to be spoken in order to convey the speaker's intent to a listener;        2. The speaker then begins to speak;        3. The listener listens to the speech and understands the flow of the conversation;        4. The listener associates the speech with the background of the speech from what the listener has experienced;        5. The listener understands the content of the speech; and        6. The listener takes an action as a result.        
If a messaging system of a computer is defined as one which simulates the message processing of a natural language and is designed for recycling by patterning a conversation in this way, a message which is closer to a natural language than in prior art message processing systems can be processed. As a result, an intelligent message system can be constructed which is more user friendly and allows a series of packets having a certain context to be exchanges with the flow of contexts being followed.